Halos at 22 deg from the sun attributed to randomly-orientated, pristine hexagonal crystals are frequently observed through ice clouds. These frequent sightings of halos formed by pristine crystals pose an apparent inconsistency with the dominance of distorted, nonpristine ice crystals indicated by in situ and remote sensing data. Furthermore, the 46 deg halo, which is associated with pristine hexagonal crystals as well, is observed far less frequently than the 22 deg halo. Considering that plausible mechanisms that could cause crystal distortion such as aggregation, sublimation, riming and collisions are stochastic processes that likely lead to distributions of crystals with varying distortion levels, here the presence of the 22 deg and 46 deg halo features in phase functions of mixtures of pristine and distorted hexagonal ice crystals is examined. We conclude that the 22 deg halo feature is generally present if the contribution by pristine crystals to the total scattering cross section is greater than only about 10% in the case of compact particles or columns, and greater than about 40% for plates. The 46 deg halo feature is present only if the mean distortion level is low and the contribution of pristine crystals to the total scattering cross section is above about 20%, 50% and 70%, in the case of compact crystals, plates and columns, respectively. These results indicate that frequent sightings of 22 deg halos are not inconsistent with the observed dominance of distorted, non-pristine ice crystals. Furthermore, the low mean distortion levels and large contributions by pristine crystals needed to produce the 461 halo features provide a potential explanation of the common sighting of the 22 deg halo without any detectable 46 deg halo.